PROJ00326

• Contract/PO #: C153498
• Contractor Name: U. S. Geological Survey
• Water District: 0
• County: Boulder
• Bill Number: SB 87-15
• Loan Projects - Doc Type: Contract Documents

ferent logging companies. The mean bulk density of
the interval was 2.23 glcm3 in hole CI and well USG::;
2.22 glcm3 in hole CIA, and 2.20 glcm3 in well A3. '
The maximum correction needed to produce an identi-

cal bulk density among the four logs was +0.03 glcm3.

In a formation that has a grain density of 2.65 glcm3, a

0.03-glcm3 correction to bulk density produces a minor
change in calculated porosity of 0.02 porosity units.
Although a small porosity error was associated
with quality control in the four logs at the test site, a
large potential error was identified by Patchett and
Coalson (1979) in a much larger sampling of wells. If
a log normalization procedure was implemented before
the log was used to estimate porosity or specific yield
some quality-control errors could be eliminated. Mos;
normalization procedures (Neinast and Knox, 1973)
require identification of calibration beds in which the
correct log response can be detennined from core ana)-
yses or by examination of many logs from nearby
wells. This technique probably is not feasible in the
Denver basin aquifer system (or other large aquifer sy~_
terns) because of the few core data available, the size
of the system (the Denver basin extends through

6,700 mi2), and the discontinuous or lenticular nature
of the lithologic units. In the absence of a rigorous no>_
malization procedure, a comparison of all density log~
from nearby wells would at least ensure that the log to
be used in porosity or specific-yield estimates is rea-
sonable and representative of conditions indicated in
other logs.

Stochastic Errors

A density log, like any nuclear log, has a stochas_
tic component in the log response caused by the ran-
dom process of radioactive decay. Instantaneous
readings from a radiation detector can vary widely.
More meaningful readings can be obtained by averag-
ing the readings over time. In geophysical logging, the
rate of movement of the logging tool up the borehole
detennines the duration of the averaging period at any
given depth. A practical compromise is needed to
achieve adequate averaging time, while still maintain-
ing a workable logging speed. The resulting density
log is not an exact measure of the radiation (and hence
bulk density) at each depth. This is indicated by a com,
parison of the density log and a 215-ft repeat section of
the log, both run in hole CIA. Differences between the
response of the two logs at I-ft depth intervals have a

mean of zero, a standard deviation of 0.03 glcm3, a

range of -0.09 to +0.11 glcm3, and a normal distribu-

tion. These statistics indicate that repeated logging of
an interval likely will produce identical mean values of
bulk density for the interval, but that density values at
specific depths can vary from log to log. Likewise, if a
density porosity log is used to estimate porosity or spe-
cific yield, log values would be better used to calculate
a mean value for a depth interval rather than to deter-
mine the porosity or specific yield at a specific depth.

Mean Grain-Density Errors

The grain density for a sandstone matrix com-

monly is set to 2.65 glcm3 in density logging, and this
value is used in equation 2 to calculate the porosity
shown on the porosity log. The mean grain density
for 139 core samples from holes CI and CIA that
have corresponding grain-size data (Robson and Banta,

1993) also is 2.65 glcm3. However, the frequency
distribution of the density data is bimodal because the
coarser grained samples have a different mean and
mode than do the finer grained samples. The mean
grain density for the coarser grained (sandstone) sam-
ples is 2.63 glcm3; the mean density for the finer

grained (mudstone) samples is 2.68 glcm3. The

0.05-glcm3 difference in mean densities produces a
minor difference in calculated porosity of 0.02 porosity
units. The range in grain density measured in the

139 core samples (2.59 to 2.79 glcm3) produces a
0.08 difference in porosity. The large range again indi-
cates the need for porosity and specific yield to be
calculated as a mean for a depth interval rather than a
value for a specific depth.

Rugosity Errors

The density log is very sensitive to the rugosity
of the borehole because any void between the logging
tool and the borehole wall is registered on the log as a
decrease in bulk density and as an increase in porosity.
Density logging tools generally are pad mounted and
decentralized to ensure close coupling between the tool
and the borehole, and density logs are run in conjunc-
tion with a caliper that measures borehole diameter.
The caliper data are used to make corrections to the
density log to compensate for changes in borehole
diameter. In some logging tools, dual detectors of dif-
ferent spacing and depths of investigation are used to
more effectively compensate for borehole rugosity.
Both means of compensation are adequate to correct
for gradual changes in diameter of a circular borehole

POROSITY ESTIMATES FROM GEOPHYSICAL LOGS 11